Voice Pathologies Classification Using GMM And SVM Classifiers

In this paper we investigate the proprieties of automatic speaker recognition (ASR) to develop a system for voice pathologies detection, where the model does not correspond to a speaker but it corresponds to group of patients who shares the same diagnostic. One of essential part in this topic is the database (described later), the samples voices (healthy and pathological) are chosen from a German database which contains many diseases, spasmodic dysphonia is proposed for this study. This problematic can be solved by statistical pattern recognition techniques where we have proposed the mel frequency cepstral coefficients (MFCC) to be modeled first, with gaussian mixture model (GMM) massively used in ASR then, they are modeled with support vector machine (SVM). The obtained results are compared in order to evaluate the more preferment classifier. The performance of each method is evaluated in a term of the accuracy, sensitivity, specificity. The best performance is obtained with 12 coefficientsMFCC, energy and second derivate along SVM with a polynomial kernel function, the classification rate is 90% for normal class and 93% for pathological class.This work is developed under MATLAB

Vocal Folds Analysis for Detection and Classification of Voice Disorder

The detection and description of pathological voice are the most important applications of voice profiling. Currently, techniques like laryngostroboscopy or surgical microlarynoscopy are popularly used for the diagnosis of voice pathologies but are invasive in nature. Disorders of vocal folds impact the quality of voice, and therefore, the accuracy of voice profiling is reduced. This paper presents a better solution to differentiate normal and pathological voices based on the glottal, physical, and acoustic and equivalent electrical parameters. These parameters have been correlated using mathematical equations and models. Results reveal that the glottal flow is strongly influenced by physical parameters like stiffness and viscosity of vocal folds in case of pathological voice. However, their direct measurement requires complex invasive medical procedures or costly and complex electronic hardware arrangements in case of non-invasive methods. Glottal parameters, on the other hand, facilitate much simpler estimation of vocal folds disorders. In this work, the authors have presented two non-invasive approaches for better accuracy and least complexity for differentiating normal and pathological voices: 1) by using correlation of glottal and physical parameters, 2)by using acoustic and equivalent electrical parameters.

Quality of Life & Phonatory and Morphological Outcomes in Cognitively Unimpaired Adolescents With Pierre Robin Sequence – A Cross-sectional Study in a 72-patient Series

We assessed the phonatory and morphological outcome of 72 cognitively unimpaired adolescents with Pierre Robin Sequence (PRS), studied their generic (Kidscreen-52), oral (COHIP-SF19) and vocal (VHI-9i) qualities of life (QoL), and sought to identify determinants of these outcomes. Two-thirds of our adolescents retained low or moderate phonation difficulties but risk factors for them were not identified. For 14%, esthetic results were considered disharmonious with no link to neonatal retrognathia severity. Only two-stage surgery appeared to have a role in improving esthetic results. The generic QoL of these adolescents was slightly lower than that of control patients, especially in dimensions concerning physical well-being, relationships and autonomy. Patients with non-isolated PRS had lower results than those with isolated PRS. Phonatory and morphological sequelae had no impact on generic QoL. Only non-isolated PRS and a low oral QoL had an impact on generic QoL. The oral QoL of these adolescents was comparable to that of control patients and significantly better than that of children with craniofacial malformations. The oral QoL of the adolescents with non-isolated PRS, was significantly worse than that of control patients and close to that of children with craniofacial malformations. The vocal QoL of our subjects was better than that of patients with other voice pathologies and better when phonation was good.

OPTIMIZATION OF DUAL-TREE COMPLEX WAVELET PACKET BASED ENTROPY FEATURES FOR VOICE PATHOLOGIES DETECTION

The Dual-Tree Complex Wavelet Packet Transform (DT-CWPT) has been successfully implemented in numerous field because it introduces limited redundancy, provides approximately shift-invariance and geometrically oriented signal in multiple dimensions where these properties are lacking in traditional wavelet transform. This paper investigates the performance of features extracted using DT-CWPT algorithms which are quantified using k-Nearest Neighbors (k-NN) and Support Vector Machine (SVM) classifiers for detecting voice pathologies. Decomposition is done on the voice signals using Shannon and Approximate entropy (ApEn) to signify the complexity of voice signals in time and frequency domain. Feature selection methods using the ReliefF algorithm and Genetic algorithm (GA) are applied to obtain the optimum features for multiclass classification. It is observed that the best accuracies obtained using DT-CWPT with ApEn entropy are 91.15 % for k-NN and 93.90 % for SVM classifiers. The proposed work provides a promising detection rate for multiple voice disorders and is useful for the development of computer-based diagnostic tools for voice pathology screening in health care facilities.

Voice Pathology Detection and Classification Using Convolutional Neural Network Model

Voice pathology disorders can be effectively detected using computer-aided voice pathology classification tools. These tools can diagnose voice pathologies at an early stage and offering appropriate treatment. This study aims to develop a powerful feature extraction voice pathology detection tool based on Deep Learning. In this paper, a pre-trained Convolutional Neural Network (CNN) was applied to a dataset of voice pathology to maximize the classification accuracy. This study also proposes a distinguished training method combined with various training strategies in order to generalize the application of the proposed system on a wide range of problems related to voice disorders. The proposed system has tested using a voice database, namely the Saarbrücken voice database (SVD). The experimental results show the proposed CNN method for speech pathology detection achieves accuracy up to 95.41%. It also obtains 94.22% and 96.13% for F1-Score and Recall. The proposed system shows a high capability of the real-clinical application that offering a fast-automatic diagnosis and treatment solutions within 3 s to achieve the classification accuracy.